br: ebs restore may cause region leaders assign on one tikv lead to all flashback request to one tikvs

[fengou1]

Bug Report

Please answer these questions before submitting your issue. Thanks!

1. Minimal reproduce step (Required)

1. create cluster, launch a tpcc to insert some data
2. wait tpcc finished, and wait all data to flush to disk. (all peer in a region has the same log term, last index, and commit index)
3. backup the cluster
4. cluster restore cluster
5. cluster ebs restore
6. deploy Grafana, and check flashback requests on tikv dashboard

2. What did you expect to see? (Required)

in condition, all peer in a region has the same log term, last index, and commit index.
flashback shall evenly send to all tikvs, not focus on some of them.

3. What did you see instead (Required)

flashback only send to one tikv or some

4. What is your TiDB version? (Required)

v6.4.0